My research interests include life cycle assessment (LCA), risk assessment (RA), energy policy, and industrial sustainability. Much of my PhD work focuses on applying advanced data science and AI techniques to identify and evaluate environmental risks across products' life cycles. A summary of the projects I have been working on is provided below.
Embedded Files
Project 1: Integration and Evaluation of Unintentional Substance Release Incidents
Unintentional substance release (USR) incidents occur far more frequently than commonly perceived and pose significant environmental threats. These include not only high-profile incidents like the Deepwater Horizon spill but also numerous smaller incidents that collectively result in substantial cumulative impacts. However, historical hazardous incidents are recorded in various formats across different datasets, creating challenges for the comprehensive evaluation of USR impacts.
To address this, I developed a series of data tools for acquisition and integration, including:
A web crawler to automatically collect incident reports.
A Natural Language Processing (NLP)-based model to identify actual leakage events from incident reports.
A data pipeline to integrate USR incidents from multiple sources.
Using these tools, I built a consolidated database comprising over 300,000 USR incidents, including oil spills, natural gas leaks, and chemical releases. I evaluated USR risks by analyzing the temporal, spatial, and statistical distribution of these incidents.
Project 2 Life Cycle Assessment on Incident Response Actions
When a USR incident occurs, cleanup and restoration efforts are essential to mitigate its impacts. However, while these activities remove spilled substances from the environment, they also consume significant resources and energy, potentially leading to additional environmental effects. To systematically evaluate this trade-off, I conducted life cycle assessment (LCA) case studies on oil spill response actions (OSRC).
To build life cycle inventories (LCIs) for the materials and equipment used in OSRC—such as vehicles, vessels, and aircraft—I utilized computer vision (CV) and large language models (LLMs) to extract resource usage data from Incident Action Plans (IAPs).
Source: BSEE
Project 3 Risk-Modified LCA framework (ongoing)
The current LCA framework overlooks the impacts of unintentional substance releases (USR), leading to underestimated evaluations and potentially biased decision-making. Traditional LCA assumes that all processes operate under "normal" conditions, implying that products never enter the environment unintentionally. However, my previous research demonstrates that USR risks cannot be ignored in a product's life cycle.
This project aims to address this gap by integrating USR risks into the LCA framework, developing a risk-modified LCA (RM-LCA) framework. This enhanced methodology will allow LCA to extend beyond "normal" operational assumptions to account for "practical" scenarios, enabling more accurate and comprehensive evaluations.
Page updated
Google Sites
Report abuse